Rotation angle detection device and torque detection device

The disclosure of Japanese Patent Applications No. 2011-051638 filed on Mar. 9, 2011 including the specification, drawings and abstract, is incorporated herein by reference in its entirety.

1. Field of the Invention

The invention relates to a rotation angle detection device that detects the rotation angle of a rotor, such as a rotor of a brushless motor, and a torque detection device.

2. Description of Related Art

In order to control a brushless motor used in an electric power steering system, or the like, currents need to be conducted through stator coils in accordance with the rotation angle of a rotor. Then, there is known a rotation angle detection device that detects the rotation angle of a rotor of a brushless motor using a detection rotor that rotates with rotation of the brushless motor. Specifically, as shown in FIG. 13 , a detection rotor 101 (hereinafter, referred to as “rotor 101 ”) includes a cylindrical magnet 102 having a plurality of magnetic pole pairs that correspond to magnetic pole pairs of a rotor of a brushless motor. Two magnetic sensors 121 and 122 are arranged around the rotor 101 at a predetermined angular interval about the rotation axis of the rotor 101 . The magnetic sensors 121 and 122 output sinusoidal signals having a predetermined phase difference. The rotation angle of the rotor 101 (the rotation angle of the rotor of the brushless motor) is detected based on these two sinusoidal signals (see, for example, Japanese Patent Application Publication No. 6-109750 (JP 6-109750 A)).

In this example, the magnet 102 has five magnetic pole pairs. That is, the magnet 102 has ten magnetic poles that are arranged at equiangular intervals. The magnetic poles are arranged at angular intervals of 36° (180° in electric angle) about the rotation axis of the rotor 101 . In addition, the two magnetic sensors 121 and 122 are arranged at an angular interval of 18° (90° in electric angle) about the rotation axis of the rotor 101 .

The direction indicated by the arrow in FIG. 13 is defined as the forward rotation direction of the detection rotor 101 . Then, as the rotor 101 is rotated in the forward direction, the rotation angle of the rotor 101 increases, whereas, as the rotor 101 is rotated in the reverse direction, the rotation angle of the rotor 101 reduces. As shown in FIG. 14 , the magnetic sensors 121 and 122 respectively output sinusoidal signals V 1 and V 2 . During one period of each of the sinusoidal signals V 1 and V 2 , the rotor 101 rotates an angle (72° (360° in electric angle)) corresponding to a single magnetic pole pair.

The angular range of one rotation of the rotor 101 is divided into five sections corresponding to the five magnetic pole pairs. The angle of the rotor 101 , which is expressed on the condition that the start position of each section is 0° and the end position of each section is 360°, is termed the electric angle θe of the rotor 101 . Here, the first magnetic sensor 121 outputs the output signal V 1 (=A 1 ·sin θe) and the second magnetic sensor 122 outputs the output signal V 2 (=A 2 ·cos θe). A 1 and A 2 are amplitudes. If the amplitudes A 1 and A 2 of the respective output signals V 1 and V 2 are equal to each other, the electric angle θe of the rotor 101 may be obtained using both output signals V 1 and V 2 according to Equation 1 below.

θ e =tan −1 (sin θ e /cos θ e )

tan −1 ( V 1 /V 2)  Equation 1

The thus obtained electric angle θe is used to control the brushless motor.

In the above-described conventional rotation angle detection device, because the amplitudes of the output signals V 1 and V 2 of the respective magnetic sensors 121 and 122 fluctuate from one magnetic pole to another due to, for example, variations in magnetic force among magnetic poles, an error may occur in detecting the rotation angle of the rotor 101 . Therefore, the output signals V 1 and V 2 of the respective magnetic sensors 121 and 122 are corrected (the amplitudes are corrected) such that the amplitudes of the output signals V 1 and V 2 of the respective magnetic sensors 121 and 122 are equal to each other based on the mechanical angle of the rotor 101 , and then the electric angle θe of the rotor 101 is computed.

When magnetic force varies among the magnetic poles, amplitude correction values used to correct the amplitudes of the output signals V 1 and V 2 of the respective magnetic sensors 121 and 122 by one period or half a period in the electric angle need to be changed. Therefore, to perform such amplitude correction, it is necessary to identify the magnetic poles sensed by the magnetic sensors 121 and 122 . After the rotor 101 rotates 360° in mechanical angle, it is possible to identify the magnetic poles sensed by the magnetic sensors 121 and 122 based on the differences in peak value among the magnetic poles. Therefore, it is possible to perform amplitude correction based on the magnetic poles sensed by the magnetic sensors 121 and 122 . However, immediately after start-up of the brushless motor, it is not possible to identify the magnetic poles sensed by the magnetic sensors 121 and 122 . Therefore, it is not possible to perform amplitude correction based on the magnetic poles sensed by the magnetic sensors 121 and 122 .

Note that a torque detection device used in an electric power steering system, or the like, detects the torsional angle of a torsion bar spring that couples an input shaft to an output shaft to thereby compute a torque applied to the input shaft. The torsional angle of the torsion bar spring becomes a value corresponding to the difference between the rotation angle of the input shaft and the rotation angle of the output shaft. Then, the rotation angle of the input shaft and the rotation angle of the output shaft are detected in a similar manner to that of the conventional rotation angle detection device, and then a torque applied to the input shaft is computed based on the difference between the detected rotation angle of the input shaft and the detected rotation angle of the output shaft.

It is an object of the invention to provide a rotation angle detection device that is able to identify magnetic poles respectively sensed by magnetic sensors in an early stage immediately after a rotor starts rotation, and a torque detection device.

An aspect of the invention relates to a rotation angle detection device that detects a rotation angle of a rotor. The rotation angle detection device includes: a multi-polar magnet that rotates with rotation of the rotor, and that has at least one magnetic pole pair having a first magnetic force and at least one magnetic pole pair having a second magnetic force; a plurality of magnetic sensors that respectively output sinusoidal output signals having a phase difference with rotation of the multi-polar magnet; a peak value detecting unit that detects peak values of output signals of the respective magnetic sensors; and a magnetic pole pair identifying unit that identifies one of the magnetic pole pairs, sensed by one of the magnetic sensors, which serves as a reference magnetic sensor, based on a combination of the peak values of the output signals of the respective magnetic sensors, detected by the peak value detecting unit. The number of the magnetic pole pairs, the number of the magnetic sensors, an arrangement pattern of the at least one magnetic pole pair having the first magnetic force and the at least one magnetic pole pair having the second magnetic force and arrangement positions of the magnetic sensors are set such that the combination of the peak values of the output signals of the respective magnetic sensors differs among the magnetic pole pairs sensed by the reference magnetic sensor.

Features, advantages, and technical and industrial significance of exemplary embodiments of the invention will be described below with reference to the accompanying drawings, in which like numerals denote like elements, and wherein;

FIG. 1 is a schematic view that shows the configuration of a first embodiment in the case where the invention is applied to a rotation angle detection device that detects the rotation angle of a rotor of a brushless motor;

FIG. 2 is a schematic view that shows the configuration of a detection rotor;

FIG. 3 is a schematic view that shows the output signal waveforms of first, second and third magnetic sensors and the magnetic pole sensed by the first magnetic sensor;

FIG. 4 is a chart that shows the details of a peak value table;

FIG. 5 is a flowchart that shows the procedure of the rotation angle computing process executed by a rotation angle computing device;

FIG. 6 is a flowchart that shows the procedure of the process of setting a relative pole number in step S 3 of FIG. 5 ;

FIG. 7A to FIG. 7C are charts for illustrating the process of setting a relative pole number;

FIG. 8 is a schematic view that shows the configuration of a second embodiment in the case where the invention is applied to a torque detection device of a vehicle steering system;

FIG. 9 is a schematic view that shows the configuration of a first magnet;

FIG. 10 is a schematic view that shows the configuration of a second magnet;

FIG. 11 is a flowchart that shows the procedure of a rotation angle computing process executed by a second rotation angle computing unit;

FIG. 12 is a schematic view that schematically shows the relative positional relationship between regions I to VIII of the first magnet and groups G 1 to G 4 of the second magnet in the case where there is no torsion in a torsion bar spring;

FIG. 13 is a schematic view for illustrating a rotation angle detecting method executed by a conventional rotation angle detection device; and

FIG. 14 is a schematic view that shows the output signal waveform of a first magnetic sensor and the output signal waveform of a second magnetic sensor.

Hereinafter, embodiments of the invention will be described in detail with reference to the accompanying drawings. FIG. 1 is a schematic view that shows the configuration of a first embodiment in the case where the invention is applied to a rotation angle detection device that detects the rotation angle of a rotor of a brushless motor. The rotation angle detection device includes a detection rotor (hereinafter, simply referred to as “rotor 1 ”) that rotates with rotation of the brushless motor 10 . As shown in FIG. 2 , the rotor 1 includes a cylindrical magnet (multi-polar magnet) 2 having a plurality of magnetic pole pairs that correspond to magnetic pole pairs of the rotor of the brushless motor 10 .

The magnet 2 has eight magnetic pole pairs M 1 to M 8 . That is, the magnet 2 has sixteen magnetic poles m 1 to m 16 that are arranged at equiangular intervals. The magnetic poles m 1 to m 16 are arranged at angular intervals of 22.5° (180° in electric angle) about the rotation axis of the rotor 1 . In this embodiment, there are two types of the magnitude of magnetic force in the magnetic pole pairs, that is, a relatively large first magnetic force and a second magnetic force that is smaller than the first magnetic force. In this embodiment, among the eight magnetic pole pairs M 1 to M 8 , the magnetic forces of the first, second, third and fifth magnetic pole pairs M 1 , M 2 , M 3 and M 5 are set at the second magnetic force (weak magnetic force), and the magnetic forces of the fourth, sixth, seventh and eighth magnetic pole pairs M 4 , M 6 , M 7 and M 8 are set at the first magnetic force (strong magnetic force).

Three magnetic sensors 21 , 22 and 23 are arranged around the rotor 1 . These three magnetic sensors 21 , 22 and 23 may also be respectively referred to as the first magnetic sensor 21 , the second magnetic sensor 22 and the third magnetic sensor 23 . In FIG. 2 , the second magnetic sensor 22 is arranged at a position spaced apart by 75° (600° in electric angle) from the first magnetic sensor 21 in the clockwise direction about the rotation axis of the rotor 1 . The third magnetic sensor 23 is arranged at a position spaced apart by 105° (840° in electric angle) from the first magnetic sensor 21 in the clockwise direction about the rotation axis of the rotor 1 .

In other words, the three magnetic sensors 21 , 22 and 23 are arranged such that, when the first magnetic sensor 21 faces the position at 0° in electric angle in the first magnetic pole pair M 1 , the second magnetic sensor 22 faces the position at 240° in electric angle in the second magnetic pole pair M 2 and the third magnetic sensor 23 faces the position at 120° in electric angle in the third magnetic pole pair M 3 . Each magnetic sensor may be, for example, a magnetic sensor that includes an element of which the electrical characteristic changes under the influence of a magnetic field, such as a Hall element and a magnetoresistive element (MR element).

The direction indicated by the arrow in FIG. 2 is defined as the forward rotation direction of the rotor 1 . Then, as the rotor 1 is rotated in the forward direction, the rotation angle of the rotor 1 increases; whereas, as the rotor 1 rotates in the reverse direction, the rotation angle of the rotor 1 reduces. As shown in FIG. 3 , the magnetic sensors 21 , 22 and 23 respectively output sinusoidal signals V 1 , V 2 and V 3 . During one period of each of the sinusoidal signals V 1 , V 2 and V 3 , the rotor 1 rotates an angle (45° (360° in electric angle)) corresponding to a single magnetic pole pair.

FIG. 3 shows the output signals V 1 , V 2 and V 3 of the respective magnetic sensors 21 , 22 and 23 with respect to a rotor angle (mechanical angle) in the case where the rotation angle of the rotor 1 is set at 0° when the boundary between the magnetic pole pair M 8 and magnetic pole pair M 1 of the rotor 1 faces the first magnetic sensor 21 . In addition, FIG. 3 shows the magnetic pole pair and magnetic pole, among the magnetic pole pairs M 1 to M 8 and magnetic poles m 1 to m 16 , sensed by the first magnetic sensor 21 with respect to a rotor angle, and shows the magnitude of magnetic force of each of the magnetic pole pairs M 1 to M 8 .

The angular range of one rotation of the rotor 1 is divided into eight sections corresponding to the eight magnetic pole pairs M 1 to M 8 , and then the angle of the rotor 1 , which is expressed on the condition that the start position of each section is 0° and the end position of each section is 360°, is termed the electric angle θe of the rotor 1 . Here, the first magnetic sensor 21 outputs the output signal V 1 (=A 1 ·sin θe) for each section corresponding to one of the eight magnetic pole pairs M 1 to M 8 . In this case, the second magnetic sensor 22 outputs the output signal V 2 (=A 2 ·sin(θe+600°))=A 2 ·sin(θe+240°)) for each section corresponding to one of the eight magnetic pole pairs M 1 to M 8 . In addition, the third magnetic sensor 23 outputs the output signal V 3 (=A 3 ·sin(θe+840°)=A 3 ·sin(θe+120°)) for each section corresponding to one of the eight magnetic pole pairs M 1 to M 8 . A 1 , A 2 and A 3 respectively indicate amplitudes. Note that, the amplitudes A 1 , A 2 and A 3 change according to the magnitude of magnetic force of each of the magnetic pole pairs M 1 to M 8 .

Thus, the magnetic sensors 21 , 22 and 23 respectively output sinusoidal signals having a predetermined phase difference 120° (electric angle). The output signals V 1 , V 2 and V 3 of the respective magnetic sensors 21 , 22 and 23 may be termed the first output signal V 1 , the second output signal V 2 and the third output signal V 3 . Referring back to FIG. 1 , the output signals V 1 , V 2 and V 3 of the respective magnetic sensors 21 , 22 and 23 are input into a rotation angle computing device 20 . The rotation angle computing device 20 identifies the magnetic poles sensed by the magnetic sensors 21 , 22 and 23 based on the output signals V 1 , V 2 and V 3 of the respective magnetic sensors 21 , 22 and 23 . The rotation angle computing device 20 corrects the amplitudes of the output signals V 1 , V 2 and V 3 of the respective magnetic sensors 21 , 22 and 23 based on the identified magnetic poles sensed by the magnetic sensors 21 , 22 and 23 . Then, the rotation angle computing device 20 computes the electric angle θe and mechanical angle θm of the rotor 1 based on the output signals of which the amplitudes have been corrected.

The electric angle θe and mechanical angle θm computed by the rotation angle computing device 20 are given to a motor controller 30 . The motor controller 30 uses the electric angle θe and mechanical angle θm given from the rotation angle computing device 20 to control the brushless motor 10 . The rotation angle computing device 20 is, for example, formed of a microcomputer, and includes a central processing unit (CPU) and memories (a ROM, a RAM, a rewritable nonvolatile memory, or the like). The nonvolatile memory of the rotation angle computing device 20 stores a peak value table for each of the magnetic sensors 21 , 22 and 23 .

FIG. 4 is a schematic view that shows the details of the peak value table. The peak value table stores, for each of the magnetic pole numbers 1 to 16 of the respective magnetic poles m 1 to m 16 , the peak values (local maximum values or local minimum values) P 1 ( 1 ) to P 1 ( 16 ) of the output signal V 1 of the first magnetic sensor 21 , corresponding to the respective magnetic poles, the peak values (local maximum values or local minimum values) P 2 ( 1 ) to P 2 ( 16 ) of the output signal V 2 of the second magnetic sensor 22 , corresponding to the respective magnetic poles, and the peak values (local maximum values or local minimum values) P 3 ( 1 ) to P 3 ( 16 ) of the output signal V 3 of the third magnetic sensor 23 , corresponding to the respective magnetic poles. Note that the magnetic sensors 21 , 22 and 23 have substantially the same characteristics, so the peak values of the output signals of the magnetic sensors 21 , 22 and 23 for the same magnetic poles are substantially the same.

The peak values are stored in the peak value table, for example, before shipment of the brushless motor 10 . The peak values stored in the amplitude correction peak value table may be obtained from one-period data or may be obtained from the mean values of multi-period data. FIG. 5 is a flowchart that shows the procedure of rotation angle computing process executed by the rotation angle computing device 20 .

The rotation angle computing process shown in FIG. 5 is repeatedly executed at predetermined computation intervals. A relative number, allocated to each magnetic pole using the magnetic pole sensed by the first magnetic sensor 21 at the time of the start of rotation angle computing process as a reference magnetic pole, is defined as a relative pole number. The relative pole number of the magnetic pole sensed by the first magnetic sensor 21 (hereinafter, referred to as “first relative pole number”) is denoted by a variable r 1 , the relative pole number of the magnetic pole sensed by the second magnetic sensor 22 (hereinafter, referred to as “second relative pole number”) is denoted by a variable r 2 , and the relative pole number of the magnetic pole sensed by the third magnetic sensor 23 (hereinafter, referred to as “third relative pole number”) is denoted by a variable r 3 . Note that each of the relative pole numbers r 1 , r 2 and r 3 is an integer of from 1 to 16 the relative pole number smaller by one than 1 is 16, and the relative pole number larger by one than 16 is 1.

In this embodiment, when the magnetic pole sensed by the first magnetic sensor 21 at the time of the start of rotation angle computing process (reference magnetic pole) is a north pole, the relative pole number “1” is allocated to that magnetic pole. On the other hand, when the magnetic pole sensed by the first magnetic sensor 21 at the time of the start of rotation angle computing process (reference magnetic pole) is a south pole, the relative pole number “2” is allocated to that magnetic pole.

In addition, the pole number of the magnetic pole sensed by the first magnetic sensor 21 is denoted by q 1 , the pole number of the magnetic pole sensed by the second magnetic sensor 22 is denoted by q 2 , and the pole number of the magnetic pole sensed by the third magnetic sensor 23 is denoted by q 3 . Note that each of the pole numbers q 1 , q 2 and q 3 is an integer of from 1 to 16, the pole number smaller by one than 1 is 16, and the pole number larger by one than 16 is 1. In addition, the magnetic pole pair number of the magnetic pole pair sensed by the first magnetic sensor 21 is denoted by Q 1 . Note that the magnetic pole pair number Q 1 is an integer of from 1 to 8, the pole number smaller by one than 1 is 8, and the pole number larger by one than 8 is 1.

When the rotation angle computing process is started, the rotation angle computing device 20 reads the values of the output signals V 1 , V 2 and V 3 of the respective magnetic sensors 21 , 22 and 23 (step S 1 ). Note that the memory (for example, RAM) of the rotation angle computing device 20 stores the output signal values that have been read so far since the computation cycle that is multiple cycles before to the current computation cycle.

In addition, in this embodiment, in order to detect the peak value (local maximum value or local minimum value) of the output signal V 1 , the output signal value of which the absolute value is larger among the read values of the output signal V 1 is saved in the memory as the peak value candidate of output signal V 1 . Similarly, in order to detect the peak value (local maximum value or local minimum value) of the output signal V 2 , the output signal value of which the absolute value is larger among the read values of the output signal V 2 is saved in the memory as the peak value candidate of the output signal V 2 . Similarly, in order to detect the peak value (local maximum value or local minimum value) of the output value V 3 , the output signal value of which the absolute value is larger among the read values of the output signal V 3 is saved in the memory as the peak value candidate of the output signal V 3 . Note that, when the zero-crossing of one of the output signals has been detected, the corresponding peak value candidate is reset to zero at a predetermined timing as will be described later.

As the values of the output signals V 1 , V 2 and V 3 are read in step S 1 , the rotation angle computing device 20 determines whether the current process is the first process after the start of rotation angle computing process (step S 2 ). When the current process is the first process after the start of rotation angle computing process (YES in step S 2 ), the rotation angle computing device 20 executes the process of setting a relative pole number (step S 3 ).

FIG. 6 shows the detailed procedure of the process of setting a relative pole number. The rotation angle computing device 20 initially determines whether the value of the first output signal V 1 is larger than 0 (step S 21 ). When the value of the first output signal V 1 is larger than 0 (YES in step S 21 ), the rotation angle computing device 20 determines that the magnetic pole sensed by the first magnetic sensor 21 (reference magnetic pole) is a north pole, and then sets the first relative pole number r 1 at 1 (step S 24 ). Then, the process proceeds to step S 26 .

On the other hand, when the value of the first output signal V 1 is smaller than or equal to 0 (NO in step S 21 ), the rotation angle computing device 20 determines whether the value of the first output signal V 1 is smaller than 0 (step S 22 ). When the value of the first output signal V 1 is smaller than 0 (YES in step S 22 ), the rotation angle computing device 20 determines that the magnetic pole sensed by the first magnetic sensor 21 (reference magnetic pole) is a south pole, and then sets the first relative pole number r 1 at 2 (step S 25 ). Then, the process proceeds to step S 26 .

When it is determined in step S 22 that the value of the first output signal V 1 is larger than or equal to 0 (NO in step S 22 ), that is, when the value of the first output signal V 1 is 0, the rotation angle computing device 20 determines whether the value of the third output signal V 3 is larger than 0 in order to determine whether the rotor rotation angle (electric angle) is 0° or 180° (step S 23 ). When the value of the third output signal V 3 is larger than 0 (YES in step S 23 ), the rotation angle computing device 20 determines that the rotor rotation angle (electric angle) is 0°, and then sets the first relative pole number r 1 at 1 (step S 24 ). Then, the process proceeds to step S 26 .

On the other hand, when the value of the third output signal V 3 is smaller than or equal to 0 (NO in step S 23 ), the rotation angle computing device 20 determines that the rotor rotation angle (electric angle) is 180°, and then sets the first relative pole number r 1 at 2 (step S 25 ). Then, the process proceeds to step S 26 . Note that the rotation angle computing device 20 may determine in step S 23 whether the value of the second output signal V 2 is smaller than 0. In this case, when the value of the second output signal V 2 is smaller than 0, the rotation angle computing device 20 causes the process to proceed to step S 24 and then sets the first relative pole number r 1 at 1. On the other hand, when the value of the second output signal V 2 is larger than or equal to 0, the rotation angle computing device 20 causes the process to proceed to step S 25 and then sets the first relative pole number r 1 at 2.

In step S 26 , the rotation angle computing device 20 determines whether the condition “V 1 ≧0 and V 2 <0” or the condition “V 1 ≦0 and V 2 >0” is satisfied. When the above condition is satisfied (YES in step S 26 ), the rotation angle computing device 20 determines that the pole number of the magnetic pole sensed by the second magnetic sensor 22 is larger by three than the pole number of the magnetic pole sensed by the first magnetic sensor 21 , and sets the second relative pole number r 2 at a number that is larger by three than the first relative pole number r 1 (r 2 =r 1 +3) (step S 27 ). Then, the process proceeds to step S 29 .

On the other hand, when the condition of step S 26 is not satisfied (NO in step S 26 ), the rotation angle computing device 20 determines that the pole number of the magnetic pole sensed by the second magnetic sensor 22 is larger by four than the pole number of the magnetic pole sensed by the first magnetic sensor 21 , and then sets the second relative pole number r 2 at a number that is larger by four than the first relative pole number r 1 (r 2 =r 1 +4) (step S 28 ). Then, the process proceeds to step S 29 .

In step S 29 , the rotation angle computing device 20 determines whether the condition “V 1 >0 and V 3 ≦0” or the condition “V 1 <0 and V 3 ≧0” is satisfied. When the above condition is satisfied (YES in step S 29 ), the rotation angle computing device 20 determines that the pole number of the magnetic pole sensed by the third magnetic sensor 23 is larger by five than the pole number of the magnetic pole sensed by the first magnetic sensor 21 , and then sets the third relative pole number r 3 at a number that is larger by five than the first relative pole number r 1 (r 3 =r 1 +5) (step S 30 ). Then, the process returns to step S 12 of FIG. 5 .

On the other hand, the condition of step S 29 is not satisfied (NO in step S 29 ), the rotation angle computing device 20 determines that the pole number of the magnetic pole sensed by the third magnetic sensor 23 is larger by four than the pole number of the magnetic pole sensed by the first magnetic sensor 21 , and then sets the third relative pole number r 3 at a number that is larger by four than the first relative pole number r 1 (r 3 =r 1 +4) (step S 31 ). Then, the process returns to step S 12 of FIG. 5 .

The reason why the second relative pole number r 2 is set at a number that is larger by three than the first relative pole number r 1 (r 2 =r 1 +3) when the condition of step S 26 is satisfied and the second relative pole number r 2 is set at a number that is larger by four than the first relative pole number r 1 (r 2 =r 1 +4) when the condition of step S 26 is not satisfied will be described. In addition, the reason why the third relative pole number r 3 is set at a number that is larger by five than the first relative pole number r 1 (r 3 =r 1 +5) when the condition of step S 29 is satisfied and the third relative pole number r 3 is set at a number that is larger by four than the first relative pole number r 1 (r 3 =r 1 +4) when the condition of S 29 is not satisfied will be described.

For example, FIG. 7A to FIG. 7C schematically show the signal waveforms of the output signals V 1 , V 2 and V 3 of the respective first, second and third magnetic sensors 21 , 22 and 23 at the time when the magnetic pole pair M 1 formed of the magnetic pole m 1 and magnetic pole m 2 of the rotor 1 passes the first magnetic sensor 21 . In FIG. 7A and FIG. 7B , S 1 and S 2 indicate the regions in which the first magnetic sensor 21 senses the magnetic pole m 1 and the second magnetic sensor 22 senses the magnetic pole m 4 . S 3 indicates the region in which the first magnetic sensor 21 senses the magnetic pole m 1 and the second magnetic sensor 22 senses the magnetic pole m 5 . S 4 and S 5 indicate the regions in which the first magnetic sensor 21 senses the magnetic pole m 2 and the second magnetic sensor 22 senses the magnetic pole m 5 . S 6 indicates the region in which the first magnetic sensor 21 senses the magnetic pole m 2 and the second magnetic sensor 22 senses the magnetic pole m 6 .

That is, in the regions S 1 , S 2 , S 4 and S 5 , the pole number of the magnetic pole sensed by the second magnetic sensor 22 is larger by three than the pole number of the magnetic pole sensed by the first magnetic sensor 21 . On the other hand, in the regions S 3 and S 6 , the pole number of the magnetic pole sensed by the second magnetic sensor 22 is larger by four than the pole number of the magnetic pole sensed by the first magnetic sensor 21 . In the regions S 1 and S 2 , the values of the output signals V 1 and V 2 satisfy the first condition that V 1 ≧0 and V 2 <0. In the region S 3 , the values of the output signals V 1 and V 2 satisfy the second condition that V 1 >0 and V 2 ≧0. In the regions S 4 and S 5 , the values of the output signals V 1 and V 2 satisfy the third condition that V 1 ≦0 and V 2 >0. In the region S 6 , the values of the output signals V 1 and V 2 satisfy the fourth condition that V 1 <0 and V 2 ≦0.

Then, the rotation angle computing device 20 determines that the pole number of the magnetic pole sensed by the second magnetic sensor 22 is larger by three than the pole number of the magnetic pole sensed by the first magnetic sensor 21 when the first condition (V 1 ≧0 and V 2 <0) or the third condition (V 1 ≦0 and V 2 >0) is satisfied, and determines that the pole number of the magnetic pole sensed by the second magnetic sensor 22 is larger by four than the pole number of the magnetic pole sensed by the first magnetic sensor 21 when neither the first condition nor the third condition is satisfied.

In addition, in FIG. 7A and FIG. 7C , S 1 indicates the region in which the first magnetic sensor 21 senses the magnetic pole m 1 and the third magnetic sensor 23 senses the magnetic pole m 5 . S 2 and S 3 indicate the regions in which the first magnetic sensor 21 senses the magnetic pole m 1 and the third magnetic sensor 23 senses the magnetic pole m 6 . S 4 indicates the region in which the first magnetic sensor 21 senses the magnetic pole m 2 and the third magnetic sensor 23 senses the magnetic pole m 6 . S 5 and S 6 indicate the regions in which the first magnetic sensor 21 senses the magnetic pole m 2 and the third magnetic sensor 23 senses the magnetic pole m 7 .

That is, in the regions S 1 and S 4 , the pole number of the magnetic pole sensed by the third magnetic sensor 23 is larger by four than the pole number of the magnetic pole sensed by the first magnetic sensor 21 . On the other hand, in the regions S 2 , S 3 , S 5 and S 6 , the pole number of the magnetic pole sensed by the second magnetic sensor 22 is larger by five than the pole number of the magnetic pole sensed by the first magnetic sensor 21 . In the region S 1 , the values of the output signals V 1 and V 3 satisfy the first condition that V 1 ≧0 and V 3 >0. In the regions S 2 and S 3 , the values of the output signals V 1 and V 3 satisfy the second condition that V 1 >0 and V 3 ≦0. In the region S 4 , the values of the output signals 1 and V 3 satisfy the third condition that V 1 ≦0 and V 3 <0. In the regions S 5 and S 6 , the values of the output signals 1 and V 3 satisfy the fourth condition that V 1 <0 and V 3 ≧0.

Then, the rotation angle computing device 20 determines that the pole number of the magnetic pole sensed by the third magnetic sensor 23 is larger by five than the pole number of the magnetic pole sensed by the first magnetic sensor 21 when the second condition (V 1 >0 and V 3 ≦0) or the fourth condition (V 1 <0 and V 3 ≧0) is satisfied, and determines that the pole number of the magnetic pole sensed by the third magnetic sensor 23 is larger by four than the pole number of the magnetic pole sensed by the first magnetic sensor 21 when neither the second condition nor the fourth condition is satisfied.

Referring back to FIG. 5 , when it is determined in step S 2 that the current process is not the first process after the start of rotation angle computing process (NO in step S 2 ), the process proceeds to step S 4 . In step S 4 , the rotation angle computing device 20 determines whether the zero-crossing, at which the sign of the value of an output signal is inverted, has been detected for each of the output signals V 1 , V 2 and V 3 based on the values of the output signals V 1 , V 2 and V 3 stored in the memory. When no zero-crossing has been detected (NO in step S 4 ), the rotation angle computing device 20 causes the process to proceed to step S 12 .

When the zero-crossing of one of the output signals V 1 , V 2 and V 3 has been detected in step S 4 (YES in step S 4 ), the rotation angle computing device 20 executes the pole number identifying process of step S 10 (described later) to determine whether the magnetic poles respectively sensed by the magnetic sensors 21 , 22 and 23 have been already identified (step S 5 ). When the magnetic poles respectively detected by the magnetic sensors 21 , 22 and 23 have not been identified, the rotation angle computing device 20 executes the process of updating a relative pole number (step S 6 ). Specifically, the rotation angle computing device 20 changes the relative pole number r 1 , r 2 or r 3 that is currently set for the magnetic sensor that output the output signal for which the zero-crossing has been detected in step S 4 , to a number larger by one or a number smaller by one based on the rotation direction of the rotor 1 .

When the rotation direction of the rotor 1 is the forward direction (direction indicated by the arrow in FIG. 2 ), the rotation angle computing device 20 updates the relative pole number r 1 , r 2 or r 3 , currently set for the magnetic sensor that outputs the output signal for which the zero-crossing has been detected in step S 4 , to a number larger by one. On the other hand, when the rotation direction of the rotor 1 is the reverse direction, the rotation angle computing device 20 updates the relative pole number r 1 , r 2 or r 3 , currently set for the magnetic sensor that outputs the output signal for which the zero-crossing has been detected, to a number smaller by one. Note that, as described above, the relative pole number smaller by one than the relative pole number “1” is “16”. In addition, the relative pole number larger by one than the relative pole number “16” is “1”.

Note that the rotation direction of the rotor 1 is determined based on the immediately preceding value and current value of the output signal for which the zero-crossing has been detected and the current value of one of the other two output signals. Specifically, when the output signal for which the zero-crossing has been detected is the first output signal V 1 , it is determined that the rotation direction is the forward direction (direction indicated by the arrow in FIG. 2 ) when the condition that “the immediately preceding value of the first output signal V 1 is larger than 0, the current value of the first output signal V 1 is smaller than or equal to 0 and the current value of the third output signal V 3 is smaller than 0 (or the current value of the second output signal V 2 is larger than 0)” or the condition that “the immediately preceding value of the first output signal V 1 is smaller than 0, the current value of the first output signal V 1 is larger than or equal to 0 and the current value of the third output signal V 3 is larger than 0 (or the current value of the second output signal V 2 is smaller than 0)” is satisfied.

On the other hand, it is determined that the rotation direction is the reverse direction when the condition that “the immediately preceding value of the first output signal V 1 is larger than or equal to 0, the current value of the first output signal V 1 is smaller than 0 and the current value of the third output signal V 3 is larger than 0 (or the current value of the second output signal V 2 is smaller than 0)” or the condition that “the immediately preceding value of the first output signal V 1 is smaller than or equal to 0, the current value of the first output signal V 1 is larger than 0 and the current value of the third output signal V 3 is smaller than 0 (or the current value of the second output signal V 2 is larger than 0)” is satisfied.

When the output signal for which the zero-crossing has been detected is the second output signal V 2 , it is determined that the rotation direction is the forward direction (direction indicated by the arrow in FIG. 2 ) when the condition that “the immediately preceding value of the second output signal V 2 is larger than 0, the current value of the second output signal V 2 is smaller than or equal to 0 and the current value of the first output signal V 1 is smaller than 0 (or the current value of the third output signal V 3 is larger than 0)” or the condition that “the immediately preceding value of the second output signal V 2 is smaller than 0, the current value of the second output signal V 2 is larger than or equal to 0 and the current value of the first output signal V 1 is larger than 0 (or the current value of the third output signal V 3 is smaller than 0)” is satisfied. On the other hand, it is determined that the rotation direction is the reverse direction when the condition that “the immediately preceding value of the second output signal V 2 is larger than or equal to 0, the current value of the second output signal V 2 is smaller than 0 and the current value of the first output signal V 1 is larger than 0 (or the current value of the third output signal V 3 is smaller than 0)” or the condition that “the immediately preceding value of the second output signal V 2 is smaller than or equal to 0, the current value of the second output signal V 2 is larger than 0 and the current value of the first output signal V 1 is smaller than 0 (or the current value of the third output signal V 3 is larger than 0)” is satisfied.

When the output signal for which the zero-crossing has been detected is the third output signal V 3 , it is determined that the rotation direction is the forward direction (direction indicated by the arrow in FIG. 2 ) when the condition that “the immediately preceding value of the third output signal V 3 is larger than 0, the current value of the third output signal V 3 is smaller than or equal to 0 and the current value of the first output signal V 1 is larger than 0 (or the current value of the second output signal V 2 is smaller than 0)” or the condition that “the immediately preceding value of the third output signal V 3 is smaller than 0, the current value of the third output signal V 3 is larger than or equal to 0 and the current value of the first output signal V 1 is smaller than 0 (or the current value of the second output signal V 2 is larger than 0)” is satisfied. On the other hand, it is determined that the rotation direction is the reverse direction when the condition that “the immediately preceding value of the third output signal V 3 is larger than or equal to 0, the current value of the third output signal V 3 is smaller than 0 and the current value of the first output signal V 1 is smaller than 0 (or the current value of the second output signal V 2 is larger than 0)” or the condition that “the immediately preceding value of the third output signal V 3 is smaller than or equal to 0, the current value of the third output signal V 3 is larger than 0 and the current value of the first output signal V 1 is larger than 0 (or the current value of the second output signal V 2 is smaller than 0)”.

When the process of updating a relative pole number in step S 6 ends, the rotation angle computing device 20 executes peak value detecting process (step S 7 ). The peak value detecting process will be specifically described. The magnetic sensor that corresponds to the output signal for which the zero-crossing has been detected in step S 4 is defined as a peak value detection magnetic sensor. The rotation angle computing device 20 initially determines whether the magnetic pole sensed by the peak value detection magnetic sensor has changed. That is, the rotation angle computing device 20 determines whether the position of the magnetic pole sensed by the peak value detection magnetic sensor at the time point when the zero-crossing of the output signal of the magnetic sensor has been detected in the immediately preceding cycle and the position of the magnetic pole sensed by the peak value detection magnetic sensor at the time point when the zero-crossing has been detected in the current cycle are different or the same. When the rotation direction of the rotor 1 is inverted, the positions of the magnetic poles at both the time points may be the same.

This determination may be, for example, made based on whether the rotation direction of the rotor 1 when the previous zero-crossing has been detected is the same as the current rotation direction of the rotor 1 . That is, when the rotation direction of the rotor 1 is the same, the rotation angle computing device 20 determines that the magnetic pole sensed by the peak value detection magnetic sensor has changed. On the other hand, when the rotation direction of the rotor 1 is different, the rotation angle computing device 20 determines that the magnetic pole sensed by the peak value detection magnetic sensor has not changed.

When it is determined that the magnetic pole sensed by the peak value detection magnetic sensor has changed, the rotation angle computing device 20 determines that a peak value has been detected, and then identifies the peak value candidate corresponding to that magnetic sensor as the peak value. On the other hand, when it is determined that the magnetic pole sensed by the peak value detection magnetic sensor has not changed, the rotation angle computing device 20 determines that no peak value has been detected. Note that the peak value includes a local maximum value that is larger than 0 and a local minimum value that is smaller than 0. In this embodiment, the magnetic pole pair sensed by the first magnetic sensor 21 is identified based on the local maximum value of the peak value. In the following description, the local maximum values of the first output signal V 1 , second output signal V 2 and third output signal V 3 detected in the peak value detecting process are respectively denoted by P 1 , P 2 and P 3 where appropriate.

When no local maximum value has been detected in the peak value detecting process (NO in step S 8 ), the rotation angle computing device 20 resets the peak value candidate, corresponding to the output signal for which the zero-crossing has been detected in step S 4 , to 0 and then causes the process to proceed to step S 12 . On the other hand, when a local maximum value has been detected in the peak value detecting process (YES in step S 8 ), the local maximum value is stored as a local maximum value P 1 , P 2 or P 3 corresponding to the output signal for which the zero-crossing has been detected in step S 4 (step S 9 ). Then, the rotation angle computing device 20 resets the peak value candidate, corresponding to the output signal for which the zero-crossing has been detected in step S 4 , to 0 and then causes the process to proceed to step S 10 .

In step S 10 , the rotation angle computing device 20 executes pole number identifying process (magnetic pole pair identifying process). The concept of the pole number identifying process will be described with reference to FIG. 3 . As shown in FIG. 3 , a threshold A (A>0) is set at the midpoint between the peak values (local maximum values) of the output signals V 1 , V 2 and V 3 for the north magnetic poles m 1 , m 3 , m 5 and m 9 in the magnetic pole pairs M 1 , M 2 , M 3 and M 5 having a weak magnetic force and the peak values (local maximum values) of the output signals V 1 , V 2 and V 3 for the north magnetic poles m 7 , m 11 , m 13 and m 15 in the magnetic pole pairs M 4 , M 6 , M 7 and M 8 having a strong magnetic force.

As is apparent from FIG. 3 , the combination of the local maximum values of the three output signals V 1 , V 2 and V 3 is different among the magnetic pole pairs M 1 to M 8 sensed by the first magnetic sensor 21 . Thus, it is possible to identify the magnetic pole pair sensed by the first magnetic sensor 21 based on the combination of the local maximum values P 1 , P 2 and P 3 of the three output signals V 1 , V 2 and V 3 . In addition, it is possible to identify the magnetic pole sensed by the first magnetic sensor 21 based on the timing at which the magnetic pole pair sensed by the first magnetic sensor 21 is identified. Then, it is possible to identify the magnetic poles respectively sensed by the second and third magnetic sensors 22 and 23 based on the magnetic pole sensed by the first magnetic sensor 21 .

The relationship between the magnetic pole pairs M 1 to M 8 sensed by the first magnetic sensor 21 and combinations of the local maximum values P 1 , P 2 and P 3 of the three output signals V 1 , V 2 and V 3 is shown in Table 1.

Note that, when the rotor 1 is rotating in the forward direction, the timing at which the magnetic pole pair sensed by the first magnetic sensor 21 is identified is the timing at which the zero-crossing of the third output signal V 3 is detected for the first time after the magnetic pole pair sensed by the first magnetic sensor 21 is changed (hereinafter, referred to as “first magnetic pole pair identification timing”). Thus, the magnetic pole pair identified based on the details of Table 1 at the first magnetic pole pair identification timing is the magnetic pole pair that has been sensed by the first magnetic sensor 21 until just before. Therefore, the magnetic pole pair sensed by the first magnetic sensor 21 at the first magnetic pole pair identification timing is the magnetic pole pair that has a magnetic pole pair number larger by one than the magnetic pole pair identified based on the details of Table 1. In addition, the magnetic pole sensed by the first magnetic sensor 21 at the first magnetic pole pair identification timing is a magnetic pole having a north pole between the two magnetic poles of the magnetic pole pair sensed by the first magnetic sensor 21 at the first magnetic pole pair identification timing.

On the other hand, when the rotor 1 is rotating in the reverse direction, the timing at which the magnetic pole pair sensed by the first magnetic sensor 21 is identified is the timing at which the zero-crossing of the first output signal V 1 is detected immediately after the magnetic pole pair sensed by the first magnetic sensor 21 is changed (hereinafter, referred to as “second magnetic pole pair identification timing”). Thus, the magnetic pole pair identified based on the details of Table 1 at the second magnetic pole pair identification timing is the magnetic pole pair that has been sensed by the first magnetic sensor 21 until just before. Therefore, the magnetic pole pair sensed by the first magnetic sensor 21 at the second magnetic pole pair identification timing is the magnetic pole pair that has a magnetic pole pair number smaller by one than the magnetic pole pair identified based on the details of Table 1. In addition, the magnetic pole sensed by the first magnetic sensor 21 at the second magnetic pole pair identification timing is a magnetic pole having a south pole between the two magnetic poles of the magnetic pole pair sensed by the first magnetic sensor 21 at the second magnetic pole pair identification timing.

Hereinafter, the pole number identifying process will be more specifically described. The pole number identifying process is executed only when one of the following first condition and second condition is satisfied in the case where the local maximum values of the three output signals V 1 , V 2 and V 3 have been already detected (stored).

First Condition The rotation direction of the rotor 1 is the forward direction and the output signal for which the zero-crossing has been detected in step S 4 is the third output signal V 3 .

Second Condition: The rotation direction of the rotor 1 is the reverse direction and the output signal for which the zero-crossing has been detected in step S 4 is the first output signal V 1 .

The first condition is used to determine whether the present time point is the first magnetic pole pair identification timing. When the first condition is satisfied, it is determined that the present time point is the first magnetic pole pair identification timing. The second condition is used to determine whether the present time point is the second magnetic pole pair identification timing. When the second condition is satisfied, it is determined that the present time point is the second magnetic pole pair identification timing.

When the first condition is satisfied, the rotation angle computing device 20 identifies the magnetic pole pair number Q 1 ′ of the magnetic pole pair, which has been sensed by the first magnetic sensor 21 until just before, based on the relationship between the magnetic pole pairs M 1 to M 8 and the combinations of the three local maximum values P 1 , P 2 and P 3 , shown in Table 1, and the combination of the detected latest local maximum values P 1 , P 2 and P 3 of the output signals V 1 , V 2 and V 3 . Then, the magnetic pole pair number that is larger by one than the magnetic pole pair number Q 1 ′ is identified as the magnetic pole pair number Q 1 (=Q 1 ′+1) of the magnetic pole pair that is currently sensed by the first magnetic sensor 21 .

Subsequently, the rotation angle computing device 20 identifies the pole number q 1 of the magnetic pole, currently sensed by the first magnetic sensor 21 , based on q 1 =2·Q 1 −1. In addition, the rotation angle computing device 20 identifies the pole number q 2 of the magnetic pole, sensed by the second magnetic sensor 22 , based on the pole number q 1 of the magnetic pole sensed by the first magnetic sensor 21 and the first and second relative pole numbers r 1 and r 2 . Specifically, the rotation angle computing device 20 identifies the pole number q 2 of the magnetic pole, sensed by the second magnetic sensor 22 , based on q 2 =(q 1 −r 1 )+r 2 .

Furthermore, the rotation angle computing device 20 identifies the pole number q 3 of the magnetic pole, sensed by the third magnetic sensor 23 , based on the pole number q 1 of the magnetic pole sensed by the first magnetic sensor 21 and the first and third relative pole numbers r 1 and r 3 . Specifically, the rotation angle computing device 20 identifies the pole number q 3 of the magnetic pole, sensed by the third magnetic sensor 23 , based on q 3 =(q 1 −r 1 )+r 3 . Then, the process proceeds to step S 12 .

On the other hand, when the second condition is satisfied, the rotation angle computing device 20 identifies the magnetic pole pair number Q 1 ′ of the magnetic pole pair, which has been sensed by the first magnetic sensor 21 until just before, based on the relationship between the magnetic pole pairs M 1 to M 8 and the combinations of the three local maximum values P 1 , P 2 and P 3 , shown in Table 1, and the combination of the detected latest local maximum values P 1 , P 2 and P 3 of the output signals V 1 , V 2 and V 3 . Then, the magnetic pole pair number that is smaller by one than the magnetic pole pair number Q 1 ′ is identified as the magnetic pole pair number Q 1 (=Q 1 ′−1) of the magnetic pole pair that is currently sensed by the first magnetic sensor 21 .

Subsequently, the rotation angle computing device 20 identifies the pole number q 1 of the magnetic pole, currently sensed by the first magnetic sensor 21 , based on q 1 =2·Q 1 . In addition, the rotation angle computing device 20 identifies the pole number q 2 of the magnetic pole, sensed by the second magnetic sensor 22 , based on q 2 =(q 1 −r 1 )+r 2 . Furthermore, the rotation angle computing device 20 identifies the pole number q 3 of the magnetic pole, sensed by the third magnetic sensor 23 , based on q 3 (q 1 −r 1 )+r 3 . Then, the process proceeds to step S 12 .

Note that, when it is determined in step S 10 that the local maximum values P 1 , P 2 and P 3 of the three output signals V 1 , V 2 and V 3 have not been detected (stored) or when it is determined that neither the first condition nor the second condition is satisfied even when the local maximum values P 1 , P 2 and P 3 of the three output signals V 1 , V 2 and V 3 have been detected, the rotation angle computing device 20 causes the process to proceed to step S 12 without executing the above described pole number identifying process.

When it is determined in step S 5 that the magnetic poles respectively sensed by the magnetic sensors 21 , 22 and 23 have been already identified (YES in step S 5 ), the rotation angle computing device 20 resets the peak value candidate corresponding to the output signal for which the zero-crossing has been detected in step S 4 to 0 and then executes the process of updating a pole number (step S 11 ). Specifically, the rotation angle computing device 20 changes the pole number q 1 , q 2 or q 3 that has been already identified for the magnetic sensor that outputs the output signal for which the zero-crossing has been detected in step S 4 , to a pole number that is larger by one or a pole number that is smaller by one based on the rotation direction of the rotor 1 .

When the rotation direction of the rotor 1 is the forward direction, the rotation angle computing device 20 updates the pole number q 1 , q 2 or q 3 that has been already identified for the magnetic sensor that outputs the output signal for which the zero-crossing has been detected, to a pole number larger by one. On the other hand, when the rotation direction of the rotor 1 is the reverse direction, the rotation angle computing device 20 updates the pole number q 1 , q 2 or q 3 that has been already identified for the magnetic sensor that outputs the output signal for which the zero-crossing has been detected, to a pole number smaller by one. Note that, the pole number smaller by one than the pole number “1” is “16”. In addition, the pole number larger by one than the pole number “16” is “1”. When the process of updating a pole number in step S 11 ends, the rotation angle computing device 20 causes the process to proceed to step S 12 .

In step S 12 , the rotation angle computing device 20 executes the amplitude correction process. That is, when the pole numbers q 1 , q 2 and q 3 of the magnetic poles respectively sensed by the magnetic sensors 21 , 22 and 23 have been identified, the rotation angle computing device 20 corrects the amplitudes of the output signals V 1 , V 2 and V 3 of the respective magnetic sensors 21 , 22 and 23 . On the other hand, when the pole numbers q 1 , q 2 and q 3 of the magnetic poles respectively sensed by the magnetic sensors 21 , 22 and 23 have not been identified, the rotation angle computing device 20 causes the process to proceed to step S 13 without executing the amplitude correction process.

When the pole numbers q 1 , q 2 and q 3 of the magnetic poles respectively sensed by the magnetic sensors 21 , 22 and 23 have been identified, the rotation angle computing device 20 acquires the peak value of the first output signal V 1 corresponding to the pole number q 1 of the magnetic pole sensed by the first magnetic sensor 21 , the peak value of the second output signal V 2 corresponding to the pole number q 2 of the magnetic pole sensed by the second magnetic sensor 22 and the peak value of the third output signal V 3 corresponding to the pole number q 3 of the magnetic pole sensed by the third magnetic sensor 23 , from the peak value table (see FIG. 4 ). The peak values of the first, second and third output signals V 1 , V 2 and V 3 , acquired from the peak value table, are respectively denoted by P 1 x , P 2 x and P 3 x . Then, the amplitudes of the output signals V 1 , V 2 and V 3 are corrected based on the acquired peak values (amplitude correction values) P 1 x , P 2 x and P 3 x and a preset reference amplitude φ.

Specifically, when signals obtained by correcting the amplitudes of the first, second and third output signals V 1 , V 2 and V 3 are respectively denoted by V 1 ′, V 2 ′ and V 3 ′, the corrected first, second and third output signals V 1 ′, V 2 ′ and V 3 ′ are respectively computed based on Equations 2, 3 and 4 below.

V 1′=( V 1 /P 1 x )×φ  Equation 2

V 2′=( V 2 /P 2 x )×φ  Equation 3

V 3′=( V 3 /P 3 x )×φ  Equation 4

When the amplitude correction process ends, the rotation angle computing device 20 causes the process to proceed to step S 13 . In step S 13 , the rotation angle computing device 20 computes the electric angle θe based on the output signals V 1 ′, V 2 ′ and V 3 ′ of which the amplitudes are corrected in step S 12 (corrected output signals V 1 ′, V 2 ′ and V 3 ′). However, when no amplitude correction is performed in step S 12 , the electric angle θe is computed based on the values of the output signals V 1 , V 2 and V 3 read in step S 1 .

Hereinafter, the case where the electric angle θe is computed based on the corrected output signals V 1 ′, V 2 ′ and V 3 ′ will be described. Note that, when the electric angle θe is computed based on the values of the output signals V 1 , V 2 and V 3 read in step S 1 , the values of the output signals V 1 , V 2 and V 3 are used in place of the corrected output signals V 1 ′, V 2 ′ and V 3 ′, so the description thereof is omitted.

The rotation angle computing device 20 computes the first electric angle θe 1 corresponding to the rotation angle of the rotor 1 based on the corrected first output signal V 1 ′ and the corrected third output signal V 3 ′. In addition, the rotation angle computing device 20 computes the second electric angle θe 2 corresponding to the rotation angle of the rotor 1 based on the corrected first output signal V 1 ′ and the corrected second output signal V 2 ′. In addition, the rotation angle computing device 20 computes the third electric angle θe 3 corresponding to the rotation angle of the rotor 1 based on the corrected second output signal V 2 ′ and the corrected third output signal V 3 ′. A method of computing the first, second and third electric angles θe 1 , θe 2 and θe 3 will be described later.

Then, the rotation angle computing device 20 , for example, computes the final electric angle θe based on Equation 5. That is, the rotation angle computing device 20 computes the mean value of the first, second and third electric angles θe 1 , θe 2 and θe 3 as the final electric angle θe.

θ e =(θ e 1 +θe 2 +θe 3)/3  Equation 5

Note that the rotation angle computing device 20 may compute the median value of the first, second and third electric angles θe 1 , θe 2 and θe 3 as the final electric angle θe. In addition, the rotation angle computing device 20 may exclude one of the first, second and third electric angles θe 1 , θe 2 and θe 3 , which is more deviated from the median value thereof than the other, and compute the mean value of the two values other than the excluded one as the final electric angle θe. Furthermore, the rotation angle computing device 20 may determine one of the first, second and third electric angles θe 1 , θe 2 and θe 3 as the final electric angle θe.

A method of computing the first electric angle θe 1 will be described. Because the reference amplitude is φ, the corrected first output signal is expressed by φ·sin θe, the corrected second output signal V 2 ′ is expressed by φ·sin(θe+240°), and the corrected third output signal V 3 ′ is expressed by φ·sin(θe÷120°). Here, for the sake of convenience of description, the reference amplitude φ is 1, and the corrected first output signal V 1 ′, the corrected second output signal V 2 ′ and the corrected third output signal V 3 ′ are respectively expressed by V 1 ′=sin θe, V 2 ′=sin(θe+240°) and V 3 ′=sin(θe+120°).

The rotation angle computing device 20 initially computes a signal V 13 ′ (=sin(θe+90°)=cos θe), of which the phase difference from the corrected first output signal V 1 ′ is 90°, based on the corrected first output signal V 1 ′ (=sin θe) and the corrected third output signal V 3 ′ (=sin(θe+120°). More specifically, the rotation angle computing device 20 computes the signal V 13 ′ based on Equation 6 below.

Equation 6 is derived based on the equation expanded from sin(θe+120°) by the addition theorem of trigonometric function. Then, the rotation angle computing device 20 computes the first electric angle θe 1 using the signal V 13 ′ (cos θe) and the corrected first output signal VP (=sin θe) based on Equation 7 below.

A method of computing the second electric angle θe 2 will be described. The rotation angle computing device 20 initially generates a signal V 12 ′ (=sin(θe+90°)=cos θe), of which the phase difference from the corrected first output signal V 1 ′ is 90°, based on the corrected first output signal V 1 ′ (=sin θe) and the corrected second output signal V 2 ′ (=sin(θe+240°)). More specifically, the rotation angle computing device 20 generates the signal V 12 ′ based on Equation 8 below.

Equation 8 is derived based on the equation expanded from sin(θe+240°) by the addition theorem of trigonometric function. Then, the rotation angle computing device 20 computes the second electric angle θe 2 using the signal V 12 ′ (=cos θe) and the corrected first output signal V 1 ′ (=sin θe) based on Equation 9 below.

A method of computing the third electric angle θe 3 will be described. The rotation angle computing device 20 initially computes the electric angle θe 3 ′ (=θe+120°), which is advanced by 120° with respect to the rotation angle (electric angle) θe of the rotor 1 , based on the corrected second output signal V 2 ′ and the corrected third output signal V 3 ′. Then, the third electric angle θe 3 is computed by subtracting 120° from the obtained rotation angle θe 3 ′.

The rotation angle computing device 20 generates a signal V 23 ′ (=sin(θe+120°+90°)), of which the phase difference from the corrected third output signal V 3 ′ is 90°, based on the corrected third output signal V 3 ′ (=sin(θe+120°)) and the corrected second output signal V 2 ′ (=sin(θe+240°)). Where θe′=θe+120°, the corrected third output signal V 3 ′ is expressed by the sinusoidal signal sin θe′ and the corrected second output signal V 2 ′ is expressed by the sinusoidal signal sin(θe+120°) of which the phase difference is advanced by 120° from the sinusoidal signal sin θe′, the signal V 23 ′ (=sin(θe′+90°)=cos θe′) of which the phase difference from the sinusoidal signal sin θe′ is 90° is obtained as in the case of the method of computing the first electric angle θe 1 . Specifically, the rotation angle computing device 20 generates the signal V 23 ′ based on Equation 10 below.

Subsequently, the rotation angle computing device 20 computes the rotation angle θe 3 ′ using the signal V 23 ′ (=cos θe′) and the corrected third output signal V 3 ′ (=sin θe′=sin(θe+120°)) based on Equation 11.

Then, the rotation angle computing device 20 computes the third electric angle θe 3 based on Equation 12.

θ e 3 =θe 3′−120°  Equation 12

When the electric angle θe is computed in step S 8 , the rotation angle computing device 20 computes the mechanical angle (absolute angle) θm of the rotor 1 . Specifically, the rotation angle computing device 20 computes the mechanical angle θm using the electric angle θe computed in step S 13 and the magnetic pole pair number Q 1 corresponding to the pole number q 1 sensed by the first magnetic sensor 21 based on Equation 13.

θ m={θe +( Q 1−1)×360°}/8  Equation 13

The rotation angle computing device 20 gives the electric angle θe computed in step S 13 and the mechanical angle θm computed in step S 14 to the motor controller 30 . Then, the process in the current computation cycle ends.

According to the first embodiment, the rotation angle computing device 20 is able to compute the electric angle θe and mechanical angle θm of the rotor 1 in each predetermined computation cycle and to give the computed electric angle θe and mechanical angle θm to the motor controller 30 . In addition, the rotation angle computing device 20 is able to identify the magnetic poles respectively sensed by the magnetic sensors 21 , 22 and 23 immediately after all the local maximum values of the first output signal V 1 , second output signal V 2 and third output signal V 3 have been detected after the start of rotation angle computing process. Thus, after the start of rotation angle computing process, it is possible to identify the magnetic poles respectively sensed by the magnetic sensors 21 , 22 and 23 after the rotor 1 (brushless motor) starts rotating and before one rotation of the rotor 1 is completed. That is, after the start of rotation angle computing process, it is possible to identify the magnetic poles respectively sensed by the magnetic sensors 21 , 22 and 23 in the early stage immediately after the rotor 1 starts rotating.

The first embodiment of the invention is described above; however, the invention may be modified into another embodiment. For example, in the first embodiment, the magnetic pole pair sensed by the first magnetic sensor 21 is identified based on the combination of the local maximum values of the three output signals V 1 , V 2 and V 3 ; instead, the magnetic pole pair sensed by the first magnetic sensor 21 may be identified based on the combination of the local minimum values of the three output signals V 1 , V 2 and V 3 .

In addition, the number of the magnetic pole pairs, the number of the magnetic sensors, the arrangement pattern of the magnetic pole pairs having a weak magnetic force and the magnetic pole pairs having a strong magnetic force and the arrangement positions of the magnetic sensors are not limited to the ones described in the above embodiment. That is, the number of the magnetic pole pairs, the number of the magnetic sensors, the arrangement pattern of the magnetic pole pairs having a weak magnetic force and the magnetic pole pairs having a strong magnetic force and the arrangement positions of the magnetic sensors may be set such that the combination of the peak values (local maximum values or local minimum values) of the respective magnetic sensors differs among the magnetic pole pairs sensed by one reference magnetic sensor.

Furthermore, even in the same magnetic pole pair, different magnitudes may be set for the positive and negative peak values to thereby determine the magnetic pole sensed by the first magnetic sensor. That is, determination may be made not by magnetic pole pairs but by magnetic poles. FIG. 8 is a schematic view that shows the configuration of a second embodiment in the case where the invention is applied to a torque detection device of a vehicle steering system.

The vehicle steering system includes an input shaft 51 , an output shaft 52 and a torsion bar spring (coupling shaft) 53 . The torsion bar spring 53 couples these shafts 51 and 52 . A steering member 54 , such as a steering wheel, is coupled to the input shaft 51 . An intermediate shaft (not shown) is coupled to the output shaft 52 . The torque detection device includes a first rotation angle detection device, a second rotation angle detection device and a torque computing unit 20 C. The first rotation angle detection device detects the rotation angle of the output shaft 52 . The second rotation angle detection device detects the rotation angle of the input shaft 51 . The torque computing unit 20 C computes the torque (steering torque) applied to the input shaft 51 based on the rotation angle of the output shaft 52 , detected by the first rotation angle detection device, and the rotation angle of the input shaft 51 , detected by the second rotation angle detection device.

The first rotation angle detection device includes a first magnet (multi-polar magnet) 2 A, three magnetic sensors 21 , 22 and 23 and a first rotation angle computing unit 20 A. The first magnet (multi-polar magnet) 2 A is coupled to the output shaft 52 so as to be rotatable together with the output shaft 52 . The three magnetic sensors 21 , 22 and 23 respectively output sinusoidal signals having a phase difference from with rotation of the first magnet 2 A. The first rotation angle computing unit 20 A computes the rotation angle (electric angle θeA and mechanical angle θmA) of the output shaft 52 based on the output signals V 1 , V 2 and V 3 of the three magnetic sensors 21 , 22 and 23 , and the like.

The second rotation angle detection device includes a second magnet (multi-polar magnet) 2 B, three magnetic sensors 24 , 25 and 26 and a second rotation angle computing device 20 B. The second magnet (multi-polar magnet) 2 B is coupled to the input shaft 51 so as to be rotatable together with the input shaft 51 . The three magnetic sensors 24 , 25 and 26 respectively output sinusoidal signals having a phase difference with rotation of the second magnet 2 B. The second rotation angle computing device 20 B computes the rotation angle (electric angle θeB and mechanical angle θmB) of the input shaft 51 based on the output signals V 4 , V 5 and V 6 of the three magnetic sensors 24 , 25 and 26 , or the like.

The first rotation angle computing unit 20 A, the second rotation angle computing unit 20 B and the torque computing unit 20 C are implemented by an electronic control unit (ECU) 60 that includes a microcomputer. The microcomputer includes a CPU and a memory (ROM, RAM, nonvolatile memory, or the like). The memory stores programs, and the like.

FIG. 9 is a schematic view that shows the configuration of the first magnet 2 A. The first magnet 2 A has a cylindrical shape. The first magnet 2 A has eight magnetic pole pairs M 1 to M 8 as in the case of the magnet 2 of the first embodiment. The eight magnetic pole pairs M 1 to M 8 are arranged at equiangular intervals in the circumferential direction. That is, the first magnet 2 A has sixteen magnetic poles m 1 to m 16 that are arranged at equiangular intervals. Angular regions corresponding to the magnetic pole pairs M 1 to M 8 in the first magnet 2 A may be respectively referred to as a first region I to an eighth region VIII. The angular width of each of the regions I to VIII is 45° (mechanical angle).

There are two types of the magnitude of magnetic force in the magnetic pole pairs of the first magnet 2 A, that is, a relatively large first magnetic force and a second magnetic force that is smaller than the first magnetic force. The arrangement pattern of the magnetic pole pairs having the first magnetic force and the magnetic pole pairs having the second magnetic force in the first magnet 2 A is the same as the arrangement pattern of the eight magnetic pole pairs of the magnet 2 of the first embodiment. Specifically, among the eight magnetic pole pairs M 1 to M 8 , the magnetic forces of the relative magnetic pole pair numbers M 1 , M 2 , M 3 and M 5 are set at the second magnetic force (weak magnetic force), and the magnetic forces of the relative magnetic pole pair numbers M 4 , M 6 , M 7 and M 8 are set at the first magnetic force (strong magnetic force).

The three magnetic sensors 21 , 22 and 23 are arranged around the first magnet 2 A. These three magnetic sensors 21 , 22 and 23 may be respectively referred to as the first magnetic sensor 21 , the second magnetic sensor 22 and the third magnetic sensor 23 . In FIG. 9 , the second magnetic sensor 22 is arranged at a position spaced apart by 75° (600° in electric angle) in the clockwise direction from the first magnetic sensor 21 about the rotation axis of the first magnet 2 A. The third magnetic sensor 23 is arranged at a position spaced apart by 105° (840° in electric angle) in the clockwise direction from the first magnetic sensor 21 about the rotation axis of the first magnet 2 A.

In other words, the three magnetic sensors 21 , 22 and 23 are arranged such that, when the first magnetic sensor 21 faces the position at 0° in electric angle in the first magnetic pole pair M 1 , the second magnetic sensor 22 faces the position at 240° in electric angle in the second magnetic pole pair M 2 and the third magnetic sensor 23 faces the position at 120° in electric angle in the third magnetic pole pair M 3 . The direction indicated by the arrow in FIG. 9 is defined as the forward rotation direction of the first magnet 2 A. Then, as the first magnet 2 A is rotated in the forward direction, the rotation angle of the first magnet 2 A increases; whereas, as the first magnet 2 A is rotated in the reverse direction, the rotation angle of the first magnet 2 A reduces. The magnetic sensors 21 , 22 and 23 respectively output sinusoidal signals V 1 , V 2 and V 3 . During one rotation of each of the sinusoidal signals V 1 , V 2 and V 3 , the first magnet 2 A rotates an angle (45° (360° in electric angle)) corresponding to a single magnetic pole pair.

The angular range of one rotation of the first magnet 2 A is divided into eight sections corresponding to the eight magnetic pole pairs M 1 to M 8 , and then the angle of the first magnet 2 A, which is expressed on the condition that the start position of each section is 0° and the end position of each section is 360°, is termed the electric angle θeA of the first magnet 2 A (the electric angle of the output shaft 52 ). Here, the first magnetic sensor 21 outputs the output signal V 1 (=A 1 ·sin θeA) for each section corresponding to one of the eight magnetic pole pairs M 1 to M 8 . In this case, the second magnetic sensor 22 outputs the output signal V 2 (=A 2 ·sin(θeA+600°)=A 2 ·sin(θeA+240°)) for each section corresponding to one of the eight magnetic pole pairs M 1 to M 8 . In addition, the third magnetic sensor 23 outputs the output signal V 3 (=A 3 ·sin(θeA+840°)=A 3 ·sin(θeA+120°)) for each section corresponding to one of the eight magnetic pole pairs M 1 to M 8 . A 1 , A 2 and A 3 respectively indicate amplitudes. However, the amplitudes A 1 , A 2 and A 3 change according to the magnitude of magnetic force of each of the magnetic pole pairs M 1 to M 8 .

Thus, the magnetic sensors 21 , 22 and 23 respectively output sinusoidal signals having a predetermined phase difference 120° (electric angle). The output signals of the magnetic sensors 21 , 22 and 23 may be respectively termed the first output signal V 1 , the second output signal V 2 and the third output signal V 3 . Referring back to FIG. 8 , the output signals V 1 , V 2 and V 3 of the respective magnetic sensors 21 , 22 and 23 are input into the first rotation angle computing unit 20 A. The first rotation angle computing unit 20 A executes rotation angle computing process similar to the rotation angle computing process (see FIG. 5 ) executed by the rotation angle computing device 20 according to the first embodiment to identify one of the region numbers I to VIII of the region corresponding to the magnetic pole pair sensed by the first magnetic sensor 21 and then compute the rotation angle (electric angle θeA and mechanical angle θmA) of the first magnet 2 A (output shaft 52 ).

FIG. 10 is a schematic view that shows the second magnet 2 B. The second magnet 2 B has a cylindrical shape. The second magnet 2 B has thirty-two magnetic pole pairs M 1 to M 32 that are arranged at equiangular intervals in the circumferential direction. That is, the second magnet 2 B has sixty-four magnetic poles that are arranged at equiangular intervals. Among the thirty-two magnetic pole pairs M 1 to M 32 , a magnetic pole pair group formed of M 1 to M 8 may be referred to as first group G 1 , a magnetic pole pair group formed of M 9 to M 16 may be referred to as second group G 2 , a magnetic pole pair group formed of M 17 to M 24 may be referred to as third group G 3 and a magnetic pole pair group formed of M 25 to M 32 may be referred to as fourth group G 4 .

Numbers 1 to 8 allocated to the eight magnetic pole pairs in each of the groups G 1 to G 4 are referred to as relative magnetic pole pair numbers [1] to [8]. There are two types of the magnitude of magnetic force in the magnetic pole pairs in the second magnet 2 B, that is, a relatively large first magnetic force and a second magnetic force that is smaller than the first magnetic force. The arrangement pattern of the magnetic pole pairs having the first magnetic force and the magnetic pole pairs having the second magnetic force in each of the groups G 1 to G 4 is the same as the arrangement pattern of the eight magnetic pole pairs M 1 to M 8 of the magnet 2 of the first embodiment. Specifically, among the eight magnetic pole pairs [1] to [8] in each group, the magnetic forces of the magnetic pole pairs respectively having the relative magnetic pole pair numbers [1], [2], [3] and [5] are set at the second magnetic force (weak magnetic force), and the magnetic forces of the magnetic pole pairs respectively having the relative magnetic pole pair numbers [4], [6], [7] and [8] are set at the first magnetic force (strong magnetic force).

The three magnetic sensors 24 , 25 and 26 are arranged around the second magnet 2 B. These three magnetic sensors 24 , 25 and 26 may be respectively referred to as the fourth magnetic sensor 24 , the fifth magnetic sensor 25 and the sixth magnetic sensor 26 . In FIG. 10 , the fifth magnetic sensor 25 is arranged at a position spaced apart by 18.75° (600° in electric angle) in the clockwise direction from the fourth magnetic sensor 24 about the rotation axis of the second magnet 213 . The sixth magnetic sensor 26 is arranged at a position spaced apart by 26.25° (840° in electric angle) in the clockwise direction from the fourth magnetic sensor 24 about the rotation axis of the second magnet 28 .

In other words, the three magnetic sensors 24 , 25 and 26 are arranged such that, when the fourth magnetic sensor 24 faces the position at 0° in electric angle in the first magnetic pole pair M 1 , the fifth magnetic sensor 25 faces the position at 240° in electric angle in the second magnetic pole pair M 2 and the sixth magnetic sensor 26 faces the position at 120° in electric angle in the third magnetic pole pair M 3 . The direction indicated by the arrow in FIG. 10 is defined as the forward rotation direction of the second magnet 2 B. Then, as the second magnet 2 B is rotated in the forward direction, the rotation angle of the second magnet 213 increases; whereas, as the second magnet 213 is rotated in the reverse direction, the rotation angle of the second magnet 2 B reduces. The magnetic sensors 24 , 25 and 26 respectively output sinusoidal signals V 4 , V 5 and V 6 . During one period of each of the sinusoidal signals V 4 , V 5 and V 6 , the second magnet 2 B rotates an angle (11.25° (360° in electric angle)) corresponding to a single magnetic pole pair.

The angular range of one rotation of the second magnet 2 B is divided into thirty-two sections corresponding to the thirty-two magnetic pole pairs M 1 to M 32 , and then the angle of the second magnet 2 B, which is expressed on the condition that the start position of each section is 0° and the end position of each section is 360′, is termed the electric angle θeB of the second magnet 2 B (the electric angle of the input shaft 51 ). Here, the fourth magnetic sensor 24 outputs the output signal V 4 (=A 4 ·sin θeB) for each section corresponding to one of the thirty-two magnetic pole pairs M 1 to M 32 . In this case, the fifth magnetic sensor 25 outputs the output signal V 5 (=A 5 ·sin(θeB+600°=))=A 5 ·sin(θeB+240°)) for each section corresponding to one of the thirty-two magnetic pole pairs M 1 to M 32 . In addition, the sixth magnetic sensor 26 outputs the output signal V 6 (=A 6 ·sin(θeB+840°))=A 6 ·sin(θeB+120°)) for each section corresponding to one of the thirty-two magnetic pole pairs M 1 to M 32 . A 4 , A 5 and A 6 respectively indicate amplitudes. However, the amplitudes A 4 , A 5 and A 6 change according to the magnitude of magnetic force of each of the magnetic pole pairs M 1 to M 32 .

Thus, the magnetic sensors 24 , 25 and 26 respectively output sinusoidal signals having a predetermined phase difference 120° (electric angle). The output signals V 4 , V 5 and V 6 of the magnetic sensors 24 , 25 and 26 may be respectively referred to as the fourth output signal V 4 , the fifth output signal V 5 and the sixth output signal V 6 . Referring back to FIG. 8 , the output signals V 4 , V 5 and V 6 of the magnetic sensors 24 , 25 and 26 are input into the second rotation angle computing unit 20 B. The second rotation angle computing unit 20 B executes rotation angle computing process shown in FIG. 11 to compute the rotation angle (electric angle θeB and mechanical angle θmB) of the second magnet 2 B (input shaft 51 ). Note that the peak value for each magnetic pole of the first magnet 2 A and the peak value for each magnetic pole of the second magnet 2 B are stored in the nonvolatile memory in the form of a peak value table as in the case of the first embodiment.

FIG. 11 is a flowchart that shows the procedure of rotation angle computing process executed by the second rotation angle computing unit 20 B. The rotation angle computing process shown in FIG. 11 is repeatedly executed at predetermined computation intervals. A relative number, allocated to each magnetic pole such that the magnetic pole sensed by the fourth magnetic sensor 24 at the time of the start of rotation angle computing process is set as a reference magnetic pole, is defined as a relative pole number. The relative pole number of the magnetic pole sensed by the fourth magnetic sensor 24 (hereinafter, referred to as “fourth relative pole number”) is denoted by a variable r 4 , the relative pole number of the magnetic pole sensed by the fifth magnetic sensor 25 (hereinafter, referred to as “fifth relative pole number”) is denoted by a variable r 5 , and the relative pole number of the magnetic pole sensed by the sixth magnetic sensor 26 (hereinafter, referred to as “sixth relative pole number”) is denoted by a variable r 6 . Note that each of the relative pole numbers r 4 , r 5 and r 6 is an integer of from 1 to 64, the relative pole number smaller by one than 1 is 64, and the relative pole number larger by one than 64 is 1.

In this embodiment, when the magnetic pole sensed by the fourth magnetic sensor 24 at the time of the start of rotation angle computing process (reference magnetic pole) is a north pole, the relative pole number “1” is allocated to that magnetic pole. On the other hand, when the magnetic pole sensed by the fourth magnetic sensor 24 at the time of the start of rotation angle computing process (reference magnetic pole) is a south pole, the relative pole number “2” is allocated to that magnetic pole.

In addition, the pole number of the magnetic pole sensed by the fourth magnetic sensor 24 is denoted by q 4 , the pole number of the magnetic pole sensed by the fifth magnetic sensor 25 is denoted by q 5 , and the pole number of the magnetic pole sensed by the sixth magnetic sensor 26 is denoted by q 6 . Note that each of the pole numbers q 4 , q 5 and q 6 is an integer of from 1 to 64, the pole number smaller by one than 1 is 64, and the pole number larger by one than 64 is 1. In addition, the relative magnetic pole pair number of the magnetic pole pair sensed by the fourth magnetic sensor 24 is denoted by [Q 4 ]. Note that the relative magnetic pole pair number [Q 4 ] is an integer of from 1 to 8, the pole number smaller by one than 1 is 8, and the pole number larger by one than 8 is 1. In addition, the magnetic pole pair number of the magnetic pole pair sensed by the fourth magnetic sensor 24 (absolute magnetic pole pair number) is denoted by Q 4 . Note that the absolute magnetic pole pair number Q 4 is an integer of from 1 to 32, the pole number smaller by one than 1 is 32, and the pole number larger by one than 32 is 1.

When the rotation angle computing process is started, the second rotation angle computing unit 20 B reads the output signals (output signal values) V 4 , V 5 and V 6 of the magnetic sensors 24 , 25 and 26 (step S 1 B). Note that the memory (for example, RAM) of the second rotation angle computing unit 20 B stores multiple sets of output signal values, including the output signal values read a predetermined time before to the latest output signal values read.

In addition, in this embodiment, in order to detect the peak value (local maximum value or local minimum value) of the output signal V 4 , the output signal value of which the absolute value is larger among the read values of the output signal V 4 is saved in the memory as the peak value candidate of the output signal V 4 . Similarly, in order to detect the peak value (local maximum value or local minimum value) of the output signal V 5 , the output signal value of which the absolute value is larger among the read values of the output signal V 5 is saved in the memory as the peak value candidate of the output signal V 5 . Similarly, in order to detect the peak value (local maximum value or local minimum value) of the output signal V 6 , the output signal value of which the absolute value is larger among the read values of the output signal V 6 is saved in the memory as the peak value candidate of the output signal V 6 . However, when the zero-crossing of one of the output signals has been detected, the corresponding peak value candidate is reset to zero at a predetermined timing as will be described later.

When the values of the output signals V 4 , V 5 and V 6 are read in step S 1 B, the second rotation angle computing unit 20 B determines whether the current process is the first process after the start of rotation angle computing process (step S 2 B). When the current process is the first process after the start of rotation angle computing process (YES in step S 2 B), the second rotation angle computing unit 20 B executes the process of setting a relative pole number (step S 3 B).

The process of setting a relative pole number is similar to the process of setting a relative pole number, described with reference to FIG. 6 . However, the first to third output signals V 1 , V 2 and V 3 in FIG. 6 are respectively replaced with the fourth to sixth output signals V 4 , V 5 and V 6 . In addition, the first to third relative pole numbers r 1 , r 2 and r 3 in FIG. 6 are respectively replaced with the fourth to sixth relative pole numbers r 4 , r 5 and r 6 .

When the process of setting a relative pole number in step S 3 B ends, the second rotation angle computing unit 20 B causes the process to proceed to step S 12 B. When it is determined in step S 2 B that the current process is not the first process after the start of rotation angle computing process (NO in step S 2 B), the process proceeds to step S 413 . In step S 4 B, the second rotation angle computing unit 20 B determines whether the zero-crossing, at which the sign of the value is inverted, has been detected for each of the values of the output signals V 4 , V 5 and V 6 based on the values of the output signals V 4 , V 5 and V 6 stored in the memory. When no zero-crossing has been detected (NO in step S 4 B), the second rotation angle computing unit 20 B causes the process to proceed to step S 12 B.

When the zero-crossing of one of the values of the output signals V 4 , V 5 and V 6 has been detected in step S 4 B (YES in step S 4 B), the second rotation angle computing unit 20 B executes pole number identifying process in step S 1013 (described later) to determine whether the magnetic poles respectively detected by the magnetic sensors 24 , 25 and 26 have been already identified (step S 5 B). When the magnetic poles respectively detected by the magnetic sensors 24 , 25 and 26 have not been identified, the second rotation angle computing unit 20 B executes the process of updating a relative pole number (step S 6 B). Specifically, the relative pole number r 4 , r 5 or r 6 currently set for the magnetic sensor that outputs the output signal for which the zero-crossing has been detected in step S 4 B is updated based on the rotation direction of the rotor 1 . More specifically, when the rotation direction of the second magnet 2 B is the forward direction (direction indicated by the arrow in FIG. 10 ), the second rotation angle computing unit 20 B updates the relative pole number r 4 , r 5 or r 6 , currently set for the magnetic sensor that outputs the output signal for which the zero-crossing has been detected in step S 4 B, to a number larger by one. On the other hand, when the rotation direction of the second magnet 2 B is the reverse direction, the second rotation angle computing unit 20 B updates the relative pole number r 4 , r 5 or r 6 , currently set for the magnetic sensor, to a number smaller by one.

When the process of updating a relative pole number in step S 6 B ends, the second rotation angle computing unit 20 B executes peak value detecting process (step S 7 B). The peak value detecting process will be specifically described. The magnetic sensor that corresponds to the output signal for which the zero-crossing has been detected in step S 4 B is defined as a peak value detection magnetic sensor. The second rotation angle computing unit 20 B initially determines whether the magnetic pole sensed by the peak value detection magnetic sensor has changed. That is, when the rotation direction of the second magnet 2 B at the time when the previous zero-crossing has been detected is the same as the current rotation direction of the second magnet 2 B, the second rotation angle computing unit 20 B determines that the magnetic pole sensed by the peak value detection magnetic sensor has changed. On the other hand, when the rotation direction of the second magnet 2 B at the time when the previous zero-crossing has been detected differs from the current rotation direction of the second magnet 2 B, the second rotation angle computing unit 20 B determines that the magnetic pole sensed by the peak value detection magnetic sensor has not changed.

When it is determined that the magnetic pole sensed by the peak value detection magnetic sensor has changed, the second rotation angle computing unit 20 B determines that the peak value has been detected, and then identifies a peak value candidate corresponding to the magnetic sensor as the peak value. On the other hand, when it is determined that the magnetic pole sensed by the peak value detection magnetic sensor has not changed, the second rotation angle computing unit 20 B determines that a peak value has not been detected. Note that the peak value includes a local maximum value that is larger than 0 and a local minimum value that is smaller than 0. In this embodiment, the magnetic pole pair sensed by the fourth magnetic sensor 24 is identified based on the local maximum value of the peak value. In the following description, the local maximum values of the fourth output signal V 4 , fifth output signal V 5 and sixth output signal V 6 detected in the peak value detecting process are respectively denoted by P 4 , P 5 and P 6 where appropriate.

When a local maximum value has not been detected in the peak value detecting process (NO in step S 8 B), the second rotation angle computing unit 20 B resets the peak value candidate, corresponding to the output signal for which the zero-crossing has been detected in step S 48 , to 0 and then causes the process to proceed to step S 12 B. On the other hand, when a local maximum value has been detected in the peak value detecting process (YES in step S 8 B), the local maximum value is stored as a local maximum value P 4 , P 5 or P 6 corresponding to the output signal for which the zero-crossing has been detected in step S 413 (step S 9 B). Then, the second rotation angle computing unit 20 B resets the peak value candidate, corresponding to the output signal for which the zero-crossing has been detected in step S 4 B, to 0 and then causes the process to proceed to step S 10 B.

In step S 10 B, the second rotation angle computing unit 20 B executes pole number identifying process (magnetic pole pair identifying process). Where a magnetic pole pair pattern in which the array of the eight magnetic pole pairs M 1 to M 8 in the first embodiment is expressed by the magnitude of magnetic force is set as a reference magnetic pole pair pattern, the magnetic pole pair pattern of each of the groups G 1 to G 4 in the second magnet 2 B is the same as the reference magnetic pole pair pattern. Therefore, the second rotation angle computing unit 20 B uses a method similar to the method of identifying the magnetic pole pair number Q 1 of the magnetic pole pair sensed by the first magnetic sensor 21 in step S 10 of FIG. 5 to make it possible to identify the relative magnetic pole pair number [Q 4 ] of the magnetic pole pair sensed by the fourth magnetic sensor 24 .

That is, the second rotation angle computing unit 20 B identifies the relative magnetic pole pair number [Q 4 ] of the magnetic pole pair sensed by the fourth magnetic sensor 24 based on the preset relationship between the relative magnetic pole pair numbers [1] to [8] of the magnetic pole pairs sensed by the fourth magnetic sensor 24 and the combinations of the local maximum values P 4 , P 5 and P 6 of the output signals V 4 , V 5 and V 6 , the detected latest local maximum values of the output signals V 4 , V 5 and V 6 , and the rotation direction of the second magnet 2 B (first magnetic pole pair identifying process).

Subsequently, the second rotation angle computing unit 20 B identifies the absolute magnetic pole pair number Q 4 of the magnetic pole pair sensed by the fourth magnetic sensor 24 based on the relative magnetic pole pair number [Q 4 ] of the magnetic pole pair sensed by the fourth magnetic sensor 24 and one of the region numbers I to VIII, which corresponds to the magnetic pole pair sensed by the first magnetic sensor 21 and which is identified by the first rotation angle computing unit 20 A (second magnetic pole pair identifying process).

FIG. 12 is a schematic view that schematically shows the relative positional relationship between the regions I to VIII of the first magnet 2 A and the groups G 1 to G 4 of the second magnet 2 B in the case where there is no torsion in the torsion bar spring 53 . If the relative magnetic pole pair number of the magnetic pole pair sensed by the fourth magnetic sensor 24 has been identified as [8] by the second rotation angle computing unit 20 B, the magnetic pole pair sensed by the fourth magnetic sensor 24 is one of M 8 , M 16 , M 34 and M 32 .

If the magnetic pole pair sensed by the fourth magnetic sensor 24 is M 8 in the first group G 1 , the maximum value of the absolute value of the torsional angle of the torsion bar spring 53 is presumed to be lower than or equal to 15° in mechanical angle, so the magnetic pole pair sensed by the first magnetic sensor 21 is one of the region II and the region III. If the magnetic pole pair sensed by the fourth magnetic sensor 24 is M 16 in the second group G 2 , the magnetic pole pair sensed by the first magnetic sensor 21 is one of the region IV and the region V.

If the magnetic pole pair sensed by the fourth magnetic sensor 24 is M 24 in the third group G 3 , the magnetic pole pair sensed by the first magnetic sensor 21 is one of the region VI and the region VII. If the magnetic pole pair sensed by the fourth magnetic sensor 24 is M 32 in the fourth group G 4 , the magnetic pole pair sensed by the first magnetic sensor 21 is one of the region VIII and the region I.

Thus, when the relative magnetic pole pair number of the magnetic pole pair sensed by the fourth magnetic sensor 24 has been identified as [8] by the second rotation angle computing unit 20 B, it is possible to identify the magnetic pole pair sensed by the fourth magnetic sensor 24 as one of M 8 , M 16 , M 24 and M 32 based on one of the region numbers I to VIII, which corresponds to the magnetic pole pair sensed by the first magnetic sensor 21 .

Similarly, when the relative magnetic pole pair number identified by the second rotation angle computing unit 20 B is a value other than [8], it is possible to identify the magnetic pole pair (absolute magnetic pole pair number) Q 4 sensed by the fourth magnetic sensor 24 in consideration of one of the region numbers I to VIII, which corresponds to the magnetic pole pair sensed by the first magnetic sensor 21 and which has been identified by the first rotation angle computing unit 20 A. Table 2 shows the relationship between the combination of the relative magnetic pole pair number [Q 4 ] of the magnetic pole pair sensed by the fourth magnetic sensor 24 and one of the region numbers I to VIII, which corresponds to the magnetic pole pair sensed by the first magnetic sensor 21 , and the absolute magnetic pole pair number Q 4 of the magnetic pole pair sensed by the fourth magnetic sensor 24 .

That is, the second rotation angle computing unit 20 B identifies the absolute magnetic pole pair number Q 4 of the magnetic pole pair sensed by the fourth magnetic sensor 24 based on the relative magnetic pole pair number [Q 4 ] of the magnetic pole pair sensed by the fourth magnetic sensor 24 , one of the region numbers I to VIII identified by the first rotation angle computing unit 20 B (region number corresponding to the magnetic pole pair sensed by the first magnetic sensor 21 ) and the details of Table 2.

As described in the first embodiment, when the rotation direction of the second magnet 2 B is the forward direction (direction indicated by the arrow in FIG. 10 ), the magnetic pole sensed by the fourth magnetic sensor 24 is a magnetic pole having a north pole between the magnetic poles of the magnetic pole pair sensed by the fourth magnetic sensor 24 . Thus, in this case, the second rotation angle computing unit 20 B identifies the pole number q 4 of the magnetic pole currently sensed by the fourth magnetic sensor 24 based on q 4 =2·Q 4 −1. In addition, the second rotation angle computing unit 20 B identifies the pole number q 5 of the magnetic pole sensed by the fifth magnetic sensor 25 based on q 5 =(q 4 −r 4 )+r 5 . Furthermore, the second rotation angle computing unit 20 B identifies the pole number q 6 of the magnetic pole sensed by the sixth magnetic sensor 26 based on q 6 =(q 4 −r 4 )+r 6 . Then, the process proceeds to step S 12 B.

On the other hand, when the rotation direction of the second magnet 2 B is the reverse direction, the magnetic pole sensed by the fourth magnetic sensor 24 is a magnetic pole having a south pole between the magnetic poles of the magnetic pole pair sensed by the fourth magnetic sensor 24 . Thus, in this case, the second rotation angle computing unit 20 B identifies the pole number q 4 of the magnetic pole currently sensed by the fourth magnetic sensor 24 based on q 4 =2·Q 4 . In addition, the second rotation angle computing unit 20 B identifies the pole number q 5 of the magnetic pole sensed by the fifth magnetic sensor 25 based on q 5 =(q 4 −r 4 )+r 5 . Furthermore, the second rotation angle computing unit 20 B identifies the pole number q 6 of the magnetic pole sensed by the sixth magnetic sensor 26 based on q 6 =(q 4 −r 4 )+r 6 . Then, the process proceeds to step S 12 B.

Note that, when the magnetic pole pair sensed by the fourth magnetic sensor 24 has not been identified, for example, when it is determined in step S 10 B that the local maximum values of the three output signals V 4 , V 5 and V 6 have not been detected (stored), or the like, the second rotation angle computing unit 20 B causes the process to proceed to step S 12 B without executing the above described pole number identifying process. When it is determined in step S 5 B that the magnetic poles respectively sensed by the magnetic sensors 24 , 25 and 26 have been already identified (YES in step S 5 B), the second rotation angle computing unit 20 B resets the peak value candidate, corresponding to the output signal for which the zero-crossing has been detected in step S 4 B, to 0 and then executes the process of updating a pole number (step S 11 B). Specifically, when the rotation direction of the second magnet 2 B is the forward direction, the pole number q 4 , q 5 or q 6 that has been already identified for the magnetic sensor that outputs the output signal for which the zero-crossing has been detected is updated to a pole number larger by one; whereas, when the rotation direction of the second magnet 2 B is the reverse direction, the pole number q 4 , q 5 or q 6 that has been already identified for the magnetic sensor that outputs the output signal for which the zero-crossing has been detected is updated to a pole number smaller by one. As the process of updating a pole number in step S 11 B ends, the second rotation angle computing unit 20 B causes the process to proceed to step S 12 B.

In step S 12 B, the second rotation angle computing unit 20 B executes amplitude correction process. That is, when the pole numbers q 4 , q 5 and q 6 of the magnetic poles respectively sensed by the magnetic sensors 24 , 25 and 26 have been identified, the second rotation angle computing unit 20 B corrects the amplitudes of the output signals V 4 , V 5 and V 6 of the magnetic sensors 24 , 25 and 26 . The amplitude correction process is similar to the amplitude correction process described in step S 12 of FIG. 5 , so the description thereof is omitted. On the other hand, when the pole numbers q 4 , q 5 and q 6 of the magnetic poles respectively sensed by the magnetic sensors 24 , 25 and 26 have not been identified, the second rotation angle computing unit 20 B causes the process to proceed to step S 13 B without executing the amplitude correction process.

When the amplitude correction process ends, the second rotation angle computing unit 20 B computes the electric angle θeB of the input shaft 51 based on the output signals V 4 ′, V 5 ′ and V 6 ′ of which the amplitudes have been corrected in step S 12 B (step S 1313 ). However, when no amplitude correction is performed in step S 12 B, the electric angle θeB is computed based on the values of the output signals V 4 , V 5 and V 6 read in step S 13 . The process of computing the electric angle θeB, executed in step S 13 B, is similar to the process of computing the electric angle θe, described in step S 13 of FIG. 5 , so the description thereof is omitted.

When the electric angle θeB is computed in step 13 B, the second rotation angle computing unit 20 B computes the mechanical angle (absolute angle) θmB of the input shaft 51 (step S 14 B). Specifically, the second rotation angle computing unit 20 B computes the mechanical angle θ mB using the electric angle θeB computed in step S 13 B and the magnetic pole pair number Q 4 corresponding to the pole number q 4 sensed by the fourth magnetic sensor 24 based on Equation 14 below.

θ mB={θeB +( Q 4−1)×360°}/32  Equation 14

The torque computing unit 20 C computes a steering torque T based on the mechanical angle θmA of the output shaft 52 , computed by the first rotation angle computing unit 20 A, the mechanical angle θmB of the input shaft 51 , computed by the second rotation angle computing unit 20 B, and the spring constant K of the torsion bar spring 53 . Specifically, the torque computing unit 20 C computes the steering torque T based on Equation 15 below.

T =(θ mA−θmB )× K   Equation 15

In the embodiment, it is possible to compute the steering torque T. In addition, it is possible to compute the steering angle θmB that is the rotation angle (mechanical angle) of the input shaft 51 and the rotation angle (mechanical angle) θmA of the output shaft 52 . The second magnet 2 B fixed to the input shaft 51 has thirty-two magnetic pole pairs, so, after the start of rotation angle computing process, the steering angle θ mB is computed in a considerably early stage immediately after the input shaft 51 starts rotation.

In the embodiment, as shown in FIG. 9 , the first magnet 2 A has the eight magnetic pole pairs M 1 to M 8 ; instead, the first magnet 2 A may have sixteen, twenty-four or thirty-two magnetic pole pairs. Thus, two, four or eight magnetic pole pairs are included in each of the regions I to VIII of the first magnet 2 A shown in FIG. 9 and FIG. 12 . In this case, the first to third magnetic sensors 21 , 22 and 23 are arranged such that the angular interval in electric angle between the consecutive two of the first to third magnetic sensors 21 , 22 and 23 are equal to the angular interval in electric angle between the consecutive two of the first to third magnetic sensors 21 , 22 and 23 shown in FIG. 9 .

In addition, in this case, the magnitude of magnetic force of each of the plurality of magnetic pole pairs included in the same region in the first magnet 2 A is set to the same strength. Then, the magnitudes of magnetic forces of the magnetic pole pairs in each of the regions I to VIII of the first magnet 2 A is set so as to be equal to the magnitudes of magnetic forces of the magnetic pole pairs M 1 to M 8 in a corresponding one of the regions I to VIII in the first magnet 2 A shown in FIG. 9 . For example, the magnitude of magnetic force of each magnetic pole pair included in the region I of the first magnet 2 A is equal to the magnitude of magnetic force of the magnetic pole pair M 1 of the above described first magnet 2 A.

Thus, the combination of the local maximum values of the three output signals V 1 , V 2 and V 3 differs among the regions I to VIII sensed by the first magnetic sensor 21 , so it is possible to identify the region sensed by the first magnetic sensor 21 based on a combination of the local maximum values of the three output signals V 1 , V 2 and V 3 . Various design changes may be applied the aspect of the invention within the range of the matter recited in the appended claims.